Healthy Choices in Online Delivery Apps

This trial tests a set of choice architecture interventions, applied in combination:

• Healthier defaults: Ramen dishes are defaulted to rice noodles (lower in calories).
• Swap prompts: Customers are prompted to switch to a miso broth (lower in calories) for applicable ramen dishes.
• Menu reordering: Menu items within key categories (e.g. ramen, sashimi) are reordered based on a composite health and revenue score, prioritising lower-calorie, high-selling items at the top.

Our target population for this intervention comprises users of a major UK food delivery platform ordering from a national Japanese restaurant chain. The primary objective is to encourage healthier ordering behaviour by reducing the average calorie content of meals ordered. This will be achieved through four targeted interventions that apply behavioural insights to the digital menu structure, making lower-calorie options more visible and easier to select while preserving user choice.

• Defaulting to a lower-calorie noodle in all ramen dishes, with the option to swap back: People tend to stick with pre-set options rather than actively changing them. By making a healthier noodle the default, this intervention increases the likelihood that customers will maintain the lower-calorie choice rather than opting out. Even if customers maintain consistent preferences, the change in the default noodle will inevitably lead to a reduction in calories. For example, a chicken ramen with regular noodles vs. one with rice noodles will naturally result in a lower-calorie choice. The default was applied to ramen dishes because these meals are the most calorific menu category, exhibit the widest range in calories, and are the most popular menu choice (accounting for approximately 50% of total revenue).
• Offering a swap to a lower-calorie broth for select ramen dishes, with messaging displayed after clicking on the item: For example: “Swap the broth in this ramen for our delicious Miso Broth, a Sapporo classic since 1953, and save 650 calories.” This approach uses prompted choice and gain framing to encourage customers to consciously select a lower-calorie alternative, appealing to ramen lovers (authenticity) and health-conscious users (calorie savings). As above, the healthy swap was applied to all non-vegan ramen dishes because these meals have high calorie contents and are among the most popular dishes.
• Reordering meals within each menu category to highlight lower-calorie options first: In food ordering, decisions are often made quickly and with minimal effort, making the initial options more likely to be chosen. We aim to encourage customers to purchase healthier options within a menu category (e.g., chicken ramen instead of spicy ramen).

The expected outputs from these interventions include an increased selection of lower-calorie meals, as healthier options are easier to find, appear more prominently, and in some cases, are set as defaults. Given these changes to the decision environment, we anticipate positive behavioural outcomes, with a measurable reduction in the average calorie content of meals ordered.

In the long term, this intervention aims to promote healthier ordering habits among food delivery platform users, making lower-calorie choices more accessible and habitual. If sustained, these changes could contribute to healthier dietary patterns among customers ordering from this restaurant chain.

However, several moderators and implementation factors could influence the effectiveness of this intervention. Compensatory behaviour (i.e., where customers add high-calorie items elsewhere in the order) may offset the calorie reduction from the initial meal choice. Additionally, habitual ordering patterns, frequency of app use, and engagement with the app’s layout will impact the intervention’s success. Usability and technical performance will remain critical to ensuring a seamless customer experience and maintaining intervention fidelity.

Primary outcome:

• Calories per ramen dish: Measures the impact of the default and swap interventions, which primarily affect ramen items.

Secondary outcomes:

• Calories per transaction: Assesses total caloric impact across the entire order.
• Revenue per restaurant: Evaluates business impact across locations.

Exploratory outcomes:

• Customer satisfaction ratings (overall order and ramen-specific).
• Revenue per transaction (to disaggregate changes in total revenue).
• Customer conversion rate (access to purchase), stratified by new vs returning customers.

We will collect:

• Transaction data: including order-level calorie content (based on mapped calorie estimates), item selection, and basket size.
• Customer feedback: including star ratings and written reviews submitted through the delivery platform.
• Revenue data: restaurant-level and transaction-level revenue, provided via the platform’s analytics tools.

A qualitative process evaluation will assess feasibility, acceptability, and perceived impact of the intervention through:

• Semi-structured interviews with key stakeholders from the restaurant and delivery platform involved in implementation.
• Open-text customer feedback submitted through standard post-order surveys on the platform.
  

The process evaluation will explore:

• Stakeholder reflections on the trial design and implementation.
• The perceived viability of scaling the interventions.
• Barriers and enablers to future adoption.
   Customer feedback data will be analysed to complement the quantitative findings and provide insights into customer satisfaction, complaints, and general perceptions of the changes.